46 巻 (1993-1994) 4 号 p. 439-455
Aftershock activities for two days after the 1923 Kanto earthquake is reinvestigated from the data of S-P time, maximum amplitude, and its period, which are newly measured on the records at the Gifu observatory. 34 aftershocks with magnitude M from 5.0 to 7.5 are identified on the records. According to locations of epicenters of aftershocks inferred from the S-P times, most of large events concentrated in the vicinities of western or eastern ends of the rupture area of the main shock. The first aftershock activity started in the western region immediately after the main shock at about 12:00 on the 1st of September and included two M=7 class events, whose magnitudes are newly determined from maximum amplitudes in vertical component of these records. At about 12:40, aftershocks began to occur in the eastern region. The next activity began at 13:00 in the western region again. From 17:00 on the 1st to 11:00 on the 2nd of September, the activity was quiet. The largest aftershock of M=7.5 and its related events occurred in the eastern region after this quietness. Then, the activity returned to the western region again at about 21:00. It is found that the number of M≥6 aftershocks is 16 for two days, which is meaningfully larger than that for the other M=8 class events in and around Japan.
The natural period of the vertical component of the seismograph was about 1s, while those of the horizontal components were about 5s. Therefore, the ratio γ of maximum amplitude in vertical component to those in horizontal components is used as a parameter which show a frequency characteristic of seismic waves from each aftershock. All the aftershocks in the eastern region have low-frequency characteristics, which show smaller γ values, comparing with those in the western region. We also found some aftershocks of which wave forms are similar to those of recent events occurring in the same region. These results indicate that the aftershock activities included the same type events as we can find in the recent seismicity and suggest that the locations of epicenters of the aftershocks inferred in the present study are proper.